PatEx: Pattern Oriented RDF Graphs Exploration
An increasing number of RDF datasets are available on the Web. In order to query these datasets, users must have information about their content as well as some knowledge of a query language such as SPARQL. Our goal is to facilitate the exploration of the
- PDF / 1,291,426 Bytes
- 13 Pages / 439.37 x 666.142 pts Page_size
- 28 Downloads / 190 Views
DAVID Lab., Univ. Versailles St Quentin, Versailles, France {hanane.ouksili,zoubida.kedad,stephane.lopes}@uvsq.fr 2 Department STEP, EDF R&D, Chatou, France [email protected]
Abstract. An increasing number of RDF datasets are available on the Web. In order to query these datasets, users must have information about their content as well as some knowledge of a query language such as SPARQL. Our goal is to facilitate the exploration of these datasets. We present in this paper PatEx, a system designed to explore RDF(S)/OWL data. PatEx provides two exploration strategies: theme-based exploration and keyword search, and users can interactively switch between the two. Moreover, the system allows the definition of patterns to formalize users’ requirements during the exploration process. We also present some evaluations performed on real datasets.
Keywords: RDF graph exploration search
1
·
Theme discovery
·
Keyword
Introduction
A huge volume of RDF datasets is available on the Web, enabling the design of novel intelligent applications. However, in order to use such datasets, the user needs to understand them. The exploration of RDF datasets is deemed to be a complex and burdensome task requiring a long time and some knowledge about the data. Formal query languages are powerful solutions to express users informational needs and to retrieve knowledge from an RDF dataset. However, in order to write these structured queries, users must have some information about the dataset schema. Besides, they must be familiar with a formal query language that could be used to query this data, such as SPARQL. To illustrate the problem and highlight our contributions, consider the dataset provided by the AIFB1 institute, which represents a research community including persons, organizations, publications and their relationships. Assume that the user is interested in the scientist named “Rudi Studer” and wants to have all his cooperators in the domain of the Semantic Web only. The user has 1
This work is supported by EDF and french ANR project CAIR. http://www.aifb.kit.edu/web/Hauptseite/en.
c Springer International Publishing Switzerland 2016 E. M´ etais et al. (Eds.): NLDB 2016, LNCS 9612, pp. 102–114, 2016. DOI: 10.1007/978-3-319-41754-7 9
PatEx: Pattern Oriented RDF Graphs Exploration
103
first to submit several queries and manually browse the dataset in order to collect all the relevant properties, which will be used to formulate the queries that will provide the final answer. Our proposal to address this issue is to combine keyword search and theme browsing to guide the user during the identification of the relevant information. We argue that if the data is presented as a set of themes along with the description of their content, it is easier to target the relevant resources and properties by exploring the relevant themes only. The user can perform a theme-based exploration, then, she may select the desired theme, the field of Semantic Web in our example, and perform a keyword search with the query “Studer Cooperator”. In
Data Loading...
